in ohio, obama is shown winning by five in the "times" poll. they had 8 points more democrats than republicans. and historically there were only 2 points more d's than r's. so that's 6 points off. so instead of romney losing by 5. he wins ohio by one. and in west virginia. >> they have owe become that winning by 2. but they have 8 points more democrats than republicans and historically there is one point more republic than democrat. that's off by a factor of 9. romney wins virginia by seven. >> that's pretty much what rove did on his little board. here's my question. can a pollster assume that all democrats and all republicans are going to vote for whoever is running in their party? can you assume that. >> yeah, you can. >> bill: with any certainty. >> yes, you can you get about 90% voting for their party. so you can it's the independents who split. what the "new york times" is doing is they are taking a poll that is probably initially accurate. i bet that all three of those polls show romney ahead. then they are waiting weighting up the number